| 1 |
This last week I have been taking a look at systrace, since first hearing |
| 2 |
about it on this list. For a while I have been looking for a "better chroot" |
| 3 |
for securing some daemons, and I think systrace could be it. |
| 4 |
|
| 5 |
Maintaining a systrace policy for a large daemon (I am working with Zope) is |
| 6 |
easier than for chroot, because there is no need to maintain a seperate set |
| 7 |
of files (or bind mounts) for the jail. This is particularly obvious when |
| 8 |
there are multiple instances of each daemon.... They can share one systrace |
| 9 |
policy, but a chroot solution would need seperate jails for each instance. |
| 10 |
|
| 11 |
The performance impact seems negligible, and the ability to log the use of |
| 12 |
incidental system calls is an unexpected bonus for intrusion detection. |
| 13 |
|
| 14 |
|
| 15 |
Is anyone looking at merging systrace with gentoo-sources? The 2003-03-22 |
| 16 |
patch has some easily resolved conflicts when merged with gentoo-sources, it |
| 17 |
compiles fine, but systrace doesnt function correctly when stressed. (more |
| 18 |
details available on request). |
| 19 |
|
| 20 |
-- |
| 21 |
Toby Dickenson |
| 22 |
http://www.geminidataloggers.com/people/tdickenson |
| 23 |
|
| 24 |
-- |
| 25 |
gentoo-hardened@g.o mailing list |
Archives